Equality, Priority, and Positional Goods* Harry Brighouse and Adam ...

Equality, Priority, and Positional Goods* Harry Brighouse and Adam Swift The question of whether those who think of themselves as egalitarians really do, or should, value equality has received considerable attention in recent years.1 Alternative principles have been offered as better capturing those distributive intuitions formerly known as ‘egalitarian’. Some endorse sufficiency—comparisons do not matter; what is important is that all have enough. Others favor giving priority to the worse off.2 The * For comments on, and discussions of, earlier drafts, we are grateful to two anonymous referees; John Baker, Christopher Bertram, Nir Eyal, Dan Hausman, Rahul Kumar, Richard Miller, Kieran Oberman, Francis Schrag, Andrew Williams; members of a reading party at the Chalet des Anglais; and audiences at the Nuffield Political Theory Workshop, the University of London’s Institute of Education, Queen’s University, the Madison Ethics Reading Group, the Cambridge Moral Sciences Club, and University College London. We’re especially grateful to A. J. Julius for a rich set of comments which prompted a radical rethink of the article. 1. From the now extensive literature, see esp. D. Parfit, “Equality or Priority?” and Larry Temkin, “Equality, Priority and the Levelling Down Objection,” both in The Ideal of Equality, ed. M. Clayton and A. Williams (Basingstoke: Palgrave Macmillan, 2000), 81–125 and 126–61; H. G. Frankfurt, “Equality as a Moral Ideal,” Ethics 98 (1987): 21–43, repr. in his The Importance of What We Care About (Cambridge: Cambridge University Press, 1988), 134–58; Roger Crisp, “Equality, Priority, and Compassion,” Ethics 113 (2003): 745–63; Larry Temkin, “Egalitarianism Defended,” Ethics 113 (2003): 764–82. 2. The view that the worse off, or worst off, should have priority admits of much further specification, some of which may be set out by contrasting Rawls’s difference principle with other variants of the Priority View. First, although Parfit’s seminal discussion presents the difference principle as prioritarian, the Priority View holds that benefiting a person matters more the worse off she is absolutely, whereas Rawls’s difference principle contains irreducible reference to relativities. For Rawls it is because they are worse off than others that benefits to the worse off matter more, whereas on the Priority View the value of the benefit depends only on how badly off they are in absolute terms. This creates scope for terminological confusion: some regard the ineliminability of relativities on the Rawlsian view as grounds for deeming that view ‘egalitarian’. We prefer to restrict the term ‘egalitarian’ to those, like Temkin, who believe there is value in equality that gives us reason to level down, on which construal neither the difference principle nor the Priority View are egalitarian. Second, the difference principle urges us to maximize the absolute position of the worst off, but it seems more plausible to regard the claims of the worst off as particularly weighty without their being that weighty. The Priority View, while Ethics 116 (April 2006): 471–497 䉷 2006 by The University of Chicago. All rights reserved. 0014-1704/2006/116030005$10.00

471

472

Ethics

April 2006

result has been to force egalitarianism stricto sensu onto the defensive. Even if sufficiency is rejected as failing to do justice to our distributive ideals, it looks perverse to prefer (a) an equally divided pie to (b) an unequally divided pie in which the smallest piece is bigger than the equal pieces of a. To advocate equality, it is alleged, is to endorse leveling down, confirming the suspicion that egalitarians are irrationally obsessed by relativities and motivated by envy. In the face of this critique, those who do not simply abandon their egalitarianism have denied this implication. Egalitarians can be pluralists about value. They think that there is a reason to level down—where doing so will make a distribution fairer—but recognize that there are also reasons not to. We are sympathetic to this response, but the result is to leave egalitarianism without much bite. The (non-person-affecting) reason to level down is unlikely to outweigh the reasons there are to prefer people to be better rather than worse off. Positional goods have a special significance in this context. They are goods with the property that one’s relative place in the distribution of the good affects one’s absolute position with respect to its value. The very fact that one is worse off than others with respect to a positional good means that one is worse off, in some respect, than one would be if that good were distributed equally. So while it might indeed be perverse to advocate leveling down all things considered, leveling down with respect to positional goods benefits absolutely, in some respect, those who would otherwise have less than others. It may thus be amenable to justification by appeal to prioritarian, and perhaps also sufficientarian, considerations. Where positional goods (or, to speak more precisely, goods with positional aspects) are concerned, one does not need to be an egalitarian to have reason to level down. This article suggests that there are more such goods than is commonly recognized and explores the implications of this insight for debates about distributive justice. The critique of equality that we have outlined resonates with much that goes on outside academia. In the wider political culture, it is frequently suggested that those who attend to relativities have simply failed to understand the benefits of growth; that attention should properly focus solely on the absolute position of the worst off and not at all on the extent of the inequality between her and others. Equality as such is a false ideal, since only a knave or a fool would prefer equality to the absolute improvements in well-being that are not only compatible with, but generated by, its unequal distribution. Part of our purpose in this holding that benefits are more valuable the worse off someone is, more modestly leaves open the issue of how much more valuable they are. For this last point, see Parfit, “Equality or Priority?” 101.

Brighouse and Swift

Equality, Priority, and Positional Goods

473

article is to connect the more abstruse and technical discussion around equality, priority, and sufficiency with the world of real politics and policy choices, explaining a number of ways in which one might see the force of the case against leveling down while continuing to insist on the reasons we have to favor more equal distributions of some goods. Many people’s distributive intuitions are somewhat good-specific: they care more that self-respect or health care or education or political influence be distributed equally than that money be so. Yet the philosophical and economic literature tends, quite properly, to abstract from these specifics.3 We hope that this article will begin to fill the gap between the two. Its structure is as follows. Section I sets out the view that equality with respect to positional goods may be justified on prioritarian grounds, without appeal to the value of equality stricto sensu. Section II shows how a concern for fair competitions may provide an independent basis for a similar conclusion. Section III discusses the variety of ways in which inequalities in the distribution of some goods may affect their possessors’ absolute position with respect to their value. The goods conventionally regarded as ‘positional’ can be thought of as special cases of a more general phenomenon, and some goods may be latently (but not manifestly) positional. More goods have positional aspects than is generally recognized, and there are thus nonegalitarian reasons to favor more equal distributions of more goods than is commonly acknowledged. The next three sections consider reasons not to level down even with respect to positional goods. Section IV notes that such goods typically have nonpositional aspects also. Their value is not exhausted by their positional aspect, and the reasons to favor their equal distribution do not apply to these other kinds of value. Section V observes that unequal distributions, or unfair competitions, may leave some people worse off than they would otherwise be with respect to particular goods, while also making them better off all things considered. In such cases we do not even have (all-things-considered) prioritarian reason to level down. In Section VI we consider the special case where unfairness may be justified on epistemic grounds. In Section VII we discuss the prospects for preventing inequalities with respect to some goods from having negative effects on those who have less of them. The goodspecific reasons to level down that we are discussing apply only insofar as goods have positional aspects. Rather than level down, it would be better if we could reduce (or eliminate altogether) the extent to which inequalities with respect to such goods negatively affect their value for those who have less. The thrust of Sections IV–VII is to defend inequality, even inequality with respect to positional goods, on all-things-considered prioritarian 3. For the state of the art, see the five papers in Economics and Philosophy 19 (2003).

474

Ethics

April 2006

grounds, thereby implicitly supporting those who emphasize the benefits of growth. But it is important to distinguish between (i) a distribution’s being justified all things considered in the circumstances and (ii) its being comprehensively justified. Section VIII suggests that there must be conceptual space morally to criticize the motivations that make it necessary for us to accept unfair inequality as the price of benefiting the worst off. The motivations that constitute the circumstances in which we are required to choose between fairness and the all-things-considered well-being of the worse off vary in moral quality. We suggest that those motivations are distinctively (but not exclusively) problematic where the goods in question are positional, that is, where the effect of some being better off than others is to make those others worse off, in some respects, than they would otherwise be. It is here that good-specific egalitarian intuitions are likely to reassert themselves. Section IX concludes. I. LEVELING DOWN AND POSITIONAL GOODS Positional goods, as we define them, are goods the absolute value of which, to their possessors, depends on those possessors’ place in the distribution of the good—on their relative standing with respect to the good in question. Their significance in relation to the leveling down objection to equality is plain. The objection suggests that it is perverse to be concerned about comparisons or relativities between people rather than how well off they are in absolute terms. But by their very nature positional goods bring together the two concerns—absolutes and relativities—that the leveling down objection keeps distinct. This is why the prioritarian case for permitting inequality does not apply to positional goods. The absolute value of a positional good depends precisely on how much of it one has compared to others. Allowing inequalities with respect to those goods, on the grounds that the inequalities benefit, or do no harm to, the worse off, makes little sense. As far as positional goods are concerned, the worse off are absolutely worse off—not just relatively so—simply because they have less than others. Fred Hirsch puts it as follows: “In one key sector—the positional sector—there is no such thing as levelling up. One’s reward is set by one’s position on the slope, and the slope itself prevents a levelling, from below as well as from above.”4 So, to take familiar examples, the labor market value of 4. F. Hirsch, Social Limits to Growth (Cambridge, MA: Harvard University Press, 1976), 175. In this book, the fullest discussion of positional goods and their economic significance, Hirsch distinguishes between goods that are positional because they are physically scarce (like Old Masters and natural landscapes) and those whose scarcity is social (like desirable jobs and social status). Among the socially scarce positional goods he distinguishes those that bring satisfaction precisely because they are scarce (snob value) and those that bring satisfaction as a consequence of their intrinsic character but are such that how much satisfaction they bring is influenced by how many others have or are pursuing them. Most

Brighouse and Swift

Equality, Priority, and Positional Goods

475

an individual’s having a master’s degree, rather than some lower level of educational qualification, depends on the distribution of educational qualifications among others in the market. In an adversarial legal system, what will concern the litigant is not so much the absolute quality of her legal representation as its quality relative to that of her opponent.5 These goods have a competitive aspect. They are valued, in part, instrumentally, as means to other goods, and their value as a means to the achievement of those goods is determined not by how much one has absolutely but by how much one has relative to relevant others. Leveling down with respect to a positional good improves the absolute position of the worst off with respect to the value of that good. Holders of master’s degrees would be absolutely better off, not just relatively so, in terms of their opportunities in the labor market, if others were deprived of the opportunity to achieve PhDs. Those with few resources to devote to their legal representation would be better off if their opponents were prevented from hiring more expensive and, let us suppose, better lawyers. As far as these goods are concerned, one does not have to be an egalitarian to have a reason to level down. Restricting inequality in itself improves the position of the worst off. Insofar, then, as goods have a positional aspect, prioritarians and egalitarians will agree that there is reason to distribute them equally. We cannot rely on growth—an increase in the total amount of the good available—to bring about a gradual improvement in the absolute position of the worse off. In these cases, there is no way that inequality could tend to improve the position of the worse off—with regard to the value of the good in question—for the only way to give more to some is to give less to others. Insofar as goods are positional, relative amount determines absolute value. This has nothing to do with envy or an irrational preoccupation with relativities or comparisons. And it means that prioritarians, and not just egalitarians, will have reason to level down with respect to those goods. II. THE VALUE OF FAIR COMPETITION Where a good is positional because its value is competitive, the value of a fair competition may provide a further reason to prefer equality to of our analysis focuses on the last kind of good, but much of it is relevant to the other kinds of positional good Hirsch is concerned with. See 28–50, esp. 50. Although much discussed by economists, positional goods have been somewhat neglected by philosophers. For notable exceptions, see J. Lichtenberg, “Consuming Because Others Consume,” Social Theory and Practice 22 (1996): 273–97; M. Hollis, “Education as a Positional Good,” Journal of Philosophy of Education 16 (1984): 235–44, and “Positional Goods,” Philosophy 18 (1984): 97–110. 5. A. Wertheimer, “The Equalization of Legal Resources,” Philosophy & Public Affairs 17 (1988): 303–22.

476

Ethics

April 2006

inequality, even where that means leveling down with respect to the good in question. Competitors being equal, in some relevant respects, may be essential to a competition between them being fair.6 The fact that some political causes, and some litigants, can command greater resources than others generates an unfairness in the competition for political power, or legal victory. The fact that wealthy parents can buy their children educational advantage over others, increasing their marketability and pushing them up the queue for well-rewarded and interesting jobs, unfairly tilts the playing field in their favor. The intuitively attractive idea of a fair competition can, in the case of positional goods, motivate an interest in equality without appeal to any more generalized prioritarianism. Even those who do not, in general, regard the claims of the worse off as having any special weight may worry about unfairness in the case of particular competitions.7 It may be tempting to cast this concern as a kind of good-specific prioritarianism. With respect to some goods, one might hold, benefits to the worse off are particularly important, namely, those goods where for them to have anything less would render them at a competitive disadvantage compared to others. In the case of some competitions, the interests of the worse off in being made better off give us reason to level down with respect to competitively positional goods—and one may believe that without endorsing any more general prioritarian commitment. We would suggest, however, that this consideration is more likely to be, and is more plausibly, framed as a variant of good-specific sufficientarianism. What matters is not so much that the worse off are made better off as that they are given enough of a chance in the competition in question. One’s reason for leveling down with respect to competitively positional goods—for preventing the wealthy from converting their money into educational advantage for their children, or superior legal representation, or more political advertising—may be that, when it comes to labor market opportunities, or legal or political victory, only this kind of leveling down will give all competitors enough of a chance. How much leveling down with respect to positional goods is justified by appeal to this kind of consideration will depend on how much of a chance is judged enough of a chance. In competitive contexts, it seems plausible both that only a fair chance is enough of a chance and that only an equal chance is a fair chance. In that case the (good-specific) 6. T. Scanlon, “The Diversity of Objections to Inequality,” in his The Difficulty of Tolerance (Cambridge: Cambridge University Press, 2003), 202–18, 205. 7. Here we are discussing cases where fair competitions are better than unfair ones in being better for at least some of the competitors. We thus set aside the ways in which fair competitions might be valuable for others (e.g., spectators may be better off watching a fair race than an unfair one).

Brighouse and Swift

Equality, Priority, and Positional Goods

477

sufficientarian and egalitarian cases will coincide, and there will be reason completely to level down. If one thought, however, that a less than equal chance could be enough of a chance, then less leveling down would be implied. One might, perhaps, hold that there was reason to level down to some extent, in order to give all competitors enough of a chance, without accepting that there was any reason to eliminate all inequality in chances due to the unequal distribution of positional goods. This sufficientarian thought perhaps underlies some attempts to limit, but not completely to remove, inequalities in children’s educational inputs, or people’s political or legal representation. Alternatively, and we would suggest more plausibly, such attempts may be motivated not so much by the thought that a less-than-equal chance is indeed a fair chance as by a willingness to trade off losses in terms of fairness for gains in terms of other values.8 III. MANIFEST AND LATENT POSITIONALITY There is, then, much to be said in favor of equality with respect to positional goods. Permitting inequalities in education or political influence or legal representation does not merely benefit some while leaving others as well off as they were before. The competitive features of the goods in question give them a zero-sum aspect; the mere fact that some have more worsens the absolute position of those who have less. And it is not simply that, in their case, generalized or even good-specific prioritarianism leads us toward equality. More specific intuitions appealing to the idea of a fair competition point in the same direction. Where egalitarians will have reasons to level down stricto sensu, prioritarians will have reason to level down with respect to goods where doing so benefits the worse off, and sufficientarians will have reason to do so where that is needed to bring up to sufficiency those who would not otherwise achieve it. Clearly, discussion and assessment of the case for equality with respect to particular goods must attend carefully to the currencies or metrics being discussed and assessed. By definition, leveling down with respect to a particular good cannot make anybody better off with respect to that good. (If it did, then that wouldn’t be leveling down with respect to that good.) Rather, leveling down with respect to a particular good 8. For the view that, in the case of opportunities for the development of merit, it is not reasonable to define a fair chance as an equal chance, see E. Anderson, “Rethinking Equality of Opportunity,” Theory and Research in Education 2 (2003): 99–110, 106. Anderson, however, is not conceiving those opportunities in competitive terms and would thus presumably deny that the goods instrumental to them are subject to the case for leveling down which appeals to the value of a fair competition. If, as we suggest, egalitarian thinking is particularly at home in competitive contexts, then clearly a good deal is likely to turn on which contexts are conceived in competitive terms.

478

Ethics

April 2006

may make some people better off with respect to the value of that good. Sometimes that value is helpfully formulated by making reference to a different good. Preventing some people from achieving doctorates does nothing to improve the education of those with master’s degrees. What it does is increase the competitive value of those degrees, improving the labor market chances of those who have achieved them. Capping input to political campaigns does not increase the political input of those who would otherwise be on the wrong end of the inequality. What it may do is increase their political influence.9 The poor person’s legal representation does not get any better just because her adversary is unable to hire a more expensive and superior lawyer. What gets better is her chance of winning. So judgments about the merits of leveling down with respect to particular goods necessarily involve evaluation of the relative importance of (a) those goods that are being leveled down (i.e., those goods with respect to which some are being made worse off while nobody is being made better off ) and (b) those goods that some come to have more of just because of the reduction in inequality with respect to the leveled-down goods. Doing this requires careful specification of the goods in question. But there are many goods the unequal distribution of which is absolutely bad for some people in some way, and many of these goods are not conventionally regarded as positional goods (or even as having a positional aspect). To this point we have talked only about what we will now call manifestly competitively positional goods. These are goods that wear their positional aspect on their sleeve, as it were, and are valued for that reason. But there are some goods that have competitive value but are not generally regarded as doing so. Here the competitive aspect is, to borrow terminology from Merton’s analysis of functionalism, latent rather than manifest.10 Consider health. It is widely believed that health differs from education in that the former, unlike the latter, is nonpositional. Education is frequently regarded as the paradigm positional good because the competitive value of my education, for me, depends on how well educated other people are. The value of my level of health, however, is, so the conventional argument goes, independent of how healthy other people are. Inequality with respect to health care, and social policies permitting or even encouraging such inequality (such as allowing a private market in health care alongside state provision free at the point of use), are widely regarded as less morally problematic than are the 9. David Estlund, “Political Quality,” Social Philosophy and Policy 17 (2000): 127–60, repr. in his Democracy (Oxford: Blackwell, 2002), 175–212 (page numbers are to Democracy). 10. R. K. Merton, “Manifest and Latent Functions,” in his Social Theory and Social Structure, rev. ed. (Glencoe, IL: Free Press, 1968), 73–138.

Brighouse and Swift

Equality, Priority, and Positional Goods

479

analogous inequalities and policies with respect to education. There may be plausible accounts whereby permitting a two-tiered or multitiered health system will, in fact, tend to worsen in absolute terms the health care available to, and hence the health of, those on the lower tiers, so that those who care about the absolute health of the least healthy will have reason to resist such policies. Such accounts do indeed yield important nonegalitarian arguments against permitting inequality, but the causal mechanisms that tend to lead to this deterioration at the bottom of the distribution are not our focus in this article.11 The point here is that the competitively positional aspect of health is not generally regarded as the basis of any objection to health inequality. In fact, however, one’s health does have competitive value. Fit and healthy people are, other things being equal, more likely to succeed in the competition for jobs and other scarce goods. Indeed, some social scientists have suggested that health is an important element in the complicated causal story that explains why economically successful parents tend to have economically successful children.12 Children of wealthy parents tend to be healthier than children of poor parents, and their being so helps to explain why they do better at school and in the labor market. If that is so, and health is indeed a determinant of children’s differential chances of achieving better or worse rewarded occupational positions, then health does indeed have a competitive, and hence positional, aspect. The value to me of my health does depend on how healthy others are. In the land of the blind, the one-eyed man is king. This is a case of a latent positional good. In effect, then, any good the differential distribution of which affects people’s chances of succeeding in the competition for other goods is properly conceived as having a competitively positional aspect. That aspect may be manifest, and it may be quantitatively significant. It may be latent and relatively trivial. Perhaps competitive positionality is manifest where it is significant and latent only where it is trivial. Perhaps not. In any case, research into the mechanisms that combine to produce inequalities in mobility chances between children born to unequally advantaged parents suggests that there are many such goods.13 11. Similarly, there are various reasons why abolishing expensive private schools in the United Kingdom might tend to improve the absolute quality of the education of those in the state sector (e.g., peer group effects or increased funding owing to the inclusion of the affluent); see A. Swift, How Not to Be a Hypocrite: School Choice for the Morally Perplexed Parent (London: Routledge Falmer, 2003). 12. S. Bowles and H. Gintis, “The Inheritance of Inequality,” Journal of Economic Perspectives 16 (2002): 3–30, 17. 13. For a useful assembling of current research on this issue, see S. Bowles, H. Gintis, and M. Osborne-Groves, eds., Unequal Chances: Family Background and Economic Success (Princeton, NJ: Princeton University Press, 2005).

480

Ethics

April 2006

Recognizing the existence of latently as well as manifestly competitively positional goods already extends the scope of application of nonegalitarian reasons to level down with respect to particular goods. But we can go further. There are many goods the unequal distribution of which is absolutely bad for some people, in some respect, in ways that have nothing to do with the competitive advantage they bestow on their possessors. Consider health, again, and suppose, as some epidemiologists have suggested, that inequalities of power or status or economic resources are bad for the health of those who have less of those goods.14 The claim is not the familiar one that those with less power or status or fewer resources tend to have worse health than those who have more. It is that the very fact of the inequality with respect to those goods worsens the health of those at the lower end of the distribution. In that case, we would have a health-specific prioritarian reason to object to the inequalities, a reason to level down with respect to those goods. Removing the inequalities will not make anybody better off with respect to power or status or resources. (Again, if it did, it would not be leveling down with respect to those goods.) But it will cause some people to be more healthy. If the health of those on the wrong end of the inequality that caused the ill health was extremely poor, we might have a goodspecific sufficientarian, and not merely a good-specific prioritarian, reason to object to the inequality. Some people might be made so ill by the inequality that we would regard their health as below a threshold of adequacy. Clearly, there are many examples along the same lines. It is a commonplace that people’s happiness depends, in part, on their (perceived) place in the distribution of other goods.15 Perhaps, as Scanlon suggests in the case of inequalities in material well-being, “the aim of avoiding stigmatization can in principle provide a reason for eliminating the benefits of the better off (or for wishing that they had never been created) even if these cannot be transferred to the worst-off.” Here the good that the material inequality deprives some people of has a psychological or subjective dimension (Scanlon says that “those who are much worse off will feel inferiority and shame at the way they must live”), but it does not have to be conceived that way.16 Whatever the feelings or self-perceptions of the people involved, we may value people’s ability to 14. R. Wilkinson, Unhealthy Societies: The Afflictions of Inequality (London: Routledge, 1996); Michael Marmot, The Status Syndrome (New York: Times Books, 2004). We simplify their thesis here for dialectical purposes; we shall present a more complex account in Sec. VII. 15. B. Frey and A. Stutzer, Happiness and Economics: How the Economy and Institutions Affect Well-Being (Princeton, NJ: Princeton University Press, 2002); R. Layard, Happiness: Lessons from a New Science (London: Allen Lane, 2005). 16. Scanlon, “The Diversity of Objections to Inequality,” 204.

Brighouse and Swift

Equality, Priority, and Positional Goods

481

participate in the communal life and shared practices of their society and object to material inequalities specifically because those with fewer resources are thereby excluded. Here the absolute good in question is something like “social inclusion.” Members of a poor but equal society may be better off, with respect to that good, just because of the lesser material inequality between them and others in their society, than members of more unequal societies who are better off in absolute material terms. Insofar as inequalities with respect to some goods affect the absolute position of some people with respect to other goods, it is possible to have nonegalitarian reasons to level down. They need not, of course, be reasons completely to level down. Presumably only material inequalities of a certain extent are bad for people’s health, or make people feel unhappy or inferior, or preclude social participation, in which case we are identifying reasons to level down only to that extent. But they are, nonetheless, reasons to prefer a more rather than less equal distribution of some goods, even where that means leveling down with respect to those goods, that do not depend on egalitarianism stricto sensu. The goods with respect to which the worse off should have priority, or of which all should have sufficient amounts, may themselves be sensitive to distributive issues. There are, then, two different ways in which positionality applies beyond the conventional case of manifestly competitive goods. On the one hand, there are latently competitive goods (e.g., health insofar as it affects labor market opportunities). On the other hand, there are goods the unequal distribution of which has adverse effects on those who have less for reasons that are independent of any competitive advantage that they bestow (e.g., material well-being insofar as it affects happiness, health, self-respect, or social inclusion).17 These too seem helpfully described as latently positional goods. In all these cases, we will have reasons to level down with respect to specific goods that do not appeal to the value of equality stricto sensu. IV. NONPOSITIONAL VALUE OF POSITIONAL GOODS Inequality, where it is unfair, may indeed be bad, and bad in ways that give us reason to level down. We believe that it is. But claims that goods should be distributed more rather than less equally are considerably 17. We focus on this article on the adverse effects of inequality on those who have less rather than more of the goods that are unequally distributed. Of course it may also be that inequality has adverse effects on those who have more of them. Presumably, if material inequality deprives people of the good of living in a society characterized by fraternal or solidaristic social relationships, then that loss is suffered as much by the rich as by the poor.

482

Ethics

April 2006

more weighty when they are justified on the person-affecting grounds that their more equal distribution will make some people better off, in some respects, than they would otherwise be. Where the unequal distribution of a good is bad for people, we don’t need to be egalitarians to have reason to level down with respect to that good. We have argued that more goods have positional aspects than is commonly acknowledged, thereby extending the scope for nonegalitarian reasoning in favor of leveling down beyond its conventional domain. There are, however, strong reasons not to level down even with respect to goods with positional aspects. Most obviously, positional goods, even manifestly competitive ones, may also have nonpositional value. It is true that, to a great extent, the value of a person’s education, to her, depends on where she stands in the distribution of that good. But that is not the only way in which education is valuable. It enables us to flourish in ways that have nothing to do with our competitiveness in labor markets. The educated person has a world of culture, complexity, and enjoyment opened to her, engaging in which is valuable in ways that are not competitive.18 Would a child of poor parents be better off, all things considered, if, as a result of leveling down, she knew a bit less mathematics or history but was able to compete in the labor market on something closer to equal terms with children of wealthy parents? Or would the absolute loss, to her, in the nonpositional value of her education outweigh the gain in its competitive value? We cannot assess the merit of leveling down with respect to education without making judgments of this kind. The good conventionally referred to as “education” is, then, valuable to its possessor in two different ways. It has, as we have sometimes put it, a positional and a nonpositional aspect. Alternatively, and perhaps more precisely, insofar as “education” is valued for two quite different reasons, we might think of it as conferring on its possessors two quite distinct goods or benefits: one an opportunity to compete in the labor market (the extent of which opportunity depends on relative position); the other an opportunity to enjoy the benefits of being an educated person (the extent of which opportunity is essentially nonpositional, or at least not competitively so). 18. Indeed, the intrinsic value of a person’s education may depend in a positive-sum rather than zero-sum way on the extent to which other people are educated. See M. Hollis, “Positional Goods,” Royal Institute of Philosophy Lectures 19 (1984): 97–110, who calls this “horizontal positionality.” A person’s ability to enjoy playing cricket depends on there being a range of other people who can play roughly as well as he does and is even enhanced by there being many around who are better than him but still willing to play against him. One’s enjoyment of literature is enhanced by the existence of others who are educated in the relevant traditions and depends on there being some who excel at the execution of, and others who excel at critical reflection on, the arts. And so on.

Brighouse and Swift

Equality, Priority, and Positional Goods

483

The noncompetitive benefits of education raise different distributive concerns from the instrumental benefits. There is no nonegalitarian case for leveling down with respect to those benefits. Still, as long as one’s relative amount of education also affects one’s chances of achieving other goods, it is impossible completely to disentangle the positional from the nonpositional aspects of education in the distribution of educational resources. Policy makers charged with designing an education system have to weigh the relative importance of positional and nonpositional benefits in deciding to which of the different principles governing their distribution to give more weight. V. PRIORITARIANISM WITH RESPECT TO WELL-BEING ALL THINGS CONSIDERED We have emphasized the importance of precision about the currencies or metrics with respect to which one is and is not leveling down. Judgments about the merits of leveling down, with respect to specific goods, must depend, inter alia, on judgments about the relative importance of the differing goods in question. We must assess the relative contribution of those goods to the overall well-being, or the well-being all things considered, of the people in question. But once we attend to well-being all things considered, rather than to the good-specific ways in which people may be made better off with respect to some goods by leveling down with respect to others, then the case for permitting inequalities is likely to reassert itself. Leveling down with respect to one good may make some people absolutely better off with respect to other goods. But making some people better off with respect to some goods does not necessarily leave them better off all things considered. Why does it ever make sense to focus on whether people are worse off than they need to be with respect to any particular good, rather than concentrating on whether they are as well off as they could be all things considered? Indeed, why should we favor leveling down, even where that is required by the value of fair competition, if, all things considered, unfair competitions benefit the unfairly treated? Here we echo those who call into question the lexical priority that Rawls gives his principle of equality of opportunity over the difference principle.19 Rawls’s second principle of justice says that social and economic inequalities are to be arranged so that they are both (a) to the greatest benefit of the least advantaged and (b) attached to offices and positions open to all under conditions of fair equality of opportunity. The fair equality of opportunity proviso asserted in b is lexically prior 19. R. Arneson, “Against Rawlsian Equality of Opportunity,” Philosophical Studies 93 (1999): 77–112; M. Clayton, “Rawls and Natural Aristocracy,” Croatian Journal of Philosophy 1 (2001): 239–59.

484

Ethics

April 2006

to the difference principle asserted in a. The principle implies that, whereas social and economic inequalities are to be arranged to the benefit of the least advantaged, the competition for the positions to which these advantages are attached is not to be so arranged. Their lexical ordering implies that fair equality of opportunity matters more than the benefit of the least advantaged and should not be sacrificed just to make them better off. Fair equality of opportunity requires that people who are similarly talented and effortful have similar prospects for attaining the positions regardless of social class background. Although the two elements may often be congruent, in some circumstances prioritizing fair equality of opportunity may require measures that depress growth—which we can think of as absolute increases in well-being—to the detriment of the least advantaged.20 Giving priority to fair equality of opportunity implies that it is more important for the similarly endowed and motivated to have equal chances of achieving unequally advantaged positions than for the least advantaged to enjoy the greatest possible benefits. In other words, it asserts that the all-things-considered well-being of the least advantaged can be compromised for the sake of distributing a particular good—opportunities for positions—equally among those with similar talent/effort quotients. For the point we wish to focus on here, the details of Rawls’s specification of his principle of fair equality of opportunity are not important. Nor is the fact that different goods are distributed by the different principles. We could revise the former so that it formulates a more general conception of equality of opportunity (not restricted to those similarly endowed and motivated). We could take the latter to govern an all-things-considered metric of advantage (not just income and wealth). Conceived thus generally, it is hard to see why we should prefer (a) a society in which all have fair or equal chances of achieving wellbeing to (b) one in which those chances are unequal, but those whose chances are worse nonetheless have better chances, of achieving wellbeing all things considered, than in a. Allowing some to receive a better-than-equal chance of achieving a particular desired outcome in a competition means giving others a worse-than-equal chance. That leaves the latter good-specifically worse off than they would be with equal chances, and it makes for an unfair competition. But it could nonetheless be the case that, all things considered, giving people unequal chances, and permitting an unfair com20. This follows as long as b is given any weight at all relative to—and not only lexical priority over—a within the principle. For Rawls’s own uncertainty about according it lexical priority, see John Rawls, Justice as Fairness: A Restatement (Cambridge, MA: Harvard University Press, 2001), 163.

Brighouse and Swift

Equality, Priority, and Positional Goods

485

petition, is better for the worse off. One argument for permitting elitist private education, and thereby allowing parents who can afford it to buy their children an unfairly good chance of getting a well-rewarded and interesting job, is that preventing their doing so may have damaging incentive effects—where ‘damaging’ means ‘deleterious to the interests of the worse off’. Deprived of this means of investing in their children’s well-being, they will have less interest in being productive, choose more leisure or consumption and less work, producing economic inefficiency and harming economic growth. It is likely also that some of what they would choose to invest in their children is itself productive, not only in helping them achieve well-rewarded positions but in terms of helping to produce goods that accrue to the worse off. Once we care about wellbeing, or prospects for well-being, all things considered, then the fact that some opportunity-related goods are positional seems a less weighty reason for distributing them equally. It remains true that in a competition you cannot improve one person’s chance of achieving a particular desired outcome without damaging somebody else’s. But if an unfair or unequal competition helps the worse off, all things considered, fairness considerations do not seem very weighty (and good-specific prioritarian considerations seem even less so). So far we have emphasized the need to assess the relative importance—the relative contribution to an individual’s well-being all things considered—of different goods. We do not know whether we have allthings-considered prioritarian reason to level down until we have made that assessment. But of course, when it comes to policy, prioritarian reasons to prefer some distributions to others have to be set against other reasons. Unless we give absolute priority to the well-being of the worst off, we are going to be balancing their interests against those of others. And even when we focus exclusively on the interests of the worse off, we have to recognize that who they are will change over time. It is one thing to permit an unfair inequality of opportunity on the ground that the individual suffering from the unfairness is likely thereby to be made better off all things considered. It is another to permit such an inequality where the individual disadvantaged by the unfairness does not herself benefit from it in any way. Perhaps it will take time for the inequality of opportunity to produce the benefits that justify it on prioritarian grounds, so long that those benefits will not be enjoyed by those on the wrong end of the inequality of opportunity. They will accrue only to future people, as yet unborn. Of course, serious prioritarian thinking must engage with well-being trade-offs between individuals, including different individuals who may at different times count as the “worse off,” as well as with trade-offs between different goods and their contribution to the all-things-considered well-being of individuals.

486

Ethics

April 2006

VI. JUSTIFYING INEQUALITY ON EPISTEMIC GROUNDS Competitive contexts, we have argued, yield special reasons to value fair distributions of relevant goods. Indeed, it seems to us that they yield reasons to favor equal distributions of such goods. Sometimes, however, it is not clear whether a context should indeed be conceived as competitive. Or, perhaps, some contexts have both competitive and noncompetitive aspects. Consider political input. There is, as we have said, a familiar line of argument—strands of which appear in Rawls’s insistence that the political liberties are special, among his basic liberties, in that they must be accorded “fair value”—which holds that political input has a positional aspect, making equality particularly important with respect to that good.21 Political input is, on this view, a competitive positional good. Some having more input than others reduces the value of the input of those others, so we have reason to value equality as discussed above. But in a recent paper, David Estlund disputes this because he conceives politics not as a competitive but as an epistemic enterprise. What matters is not that people have equal input but that democratic processes tend to reach decisions that are right by some independent moral standard. Although, other things being equal, equality of input does tend to improve epistemic quality, other things are not always equal, and sheer quantity of input matters also. According to Estlund: “Political egalitarianism neglects the fact that a small and limited discussion is not as valuable a guide to important practical decisions as is fuller, more extensive discussion. . . . If equal influence can only be achieved at lower levels of input, then the epistemic advantages of a wider discussion might, from any reasonable point of view, outweigh the disadvantages of some degree of unequal influence. This is not just a logical possibility but a real possibility in democratic politics.”22 Estlund advances two objections to the case for political equality. The first, which he does not pursue, appeals to the distinction between good-specific and all-thingsconsidered prioritarianism. He wonders why the political egalitarian ignores the possibility that “all might be made better off even if not all are given more political input or influence.”23 The second rejects the view that political input is a competitive good. Political influence may be. Of that it may be true to say that the more some have the less others have. But political influence is not the only thing that matters when it comes to assessing political procedures. Political input matters too and 21. See John Rawls, Political Liberalism, rev. ed. (New York: Columbia University Press, 1996), 356–63; H. Brighouse, “Political Equality in Justice as Fairness,” Philosophical Studies 86 (1997): 155–84. 22. Estlund, “Political Quality,” 179. 23. Ibid., 181.

Brighouse and Swift

Equality, Priority, and Positional Goods

487

can be epistemically valuable. Given that, why insist on political equality rather than permitting inequality of political input where it is likely to promote right answers—especially if that inequality of input actually increases the political input of those with least input? If permitting inequality of input is the only way to increase both total input and the input of the least advantaged, in ways that will promote epistemic quality—perhaps for incentive reasons analogous to those familiar in discussions of economic inequality—then on his view it should be permitted. Estlund does consider the possibility that political inequality might be epistemically superior without also increasing the political input of those with least. As he puts it, “Suppose that unequal input was not mutually beneficial as compared with an equal distribution—that it did not produce more input for all or more social goods for all—but that the unequal scheme did lead more reliably to just decisions. Just decisions will not necessarily benefit everyone. . . . But whether or not it is mutually beneficial, the greater substantive justice of the outcomes of the unequal distribution looks like a powerful reason in its favour.”24 This is surely right. We suggested in note 7 above that a spectator may benefit from the fact that a sports competition is fair, quite independently of any benefit accruing to the participants. But those subject to a political decision may benefit from the fact that the decision-making processes of her polity are organized so as to maximize their epistemic value, even where that makes those processes unfair. Here, then, is a scenario in which the context has both a competitive and a noncompetitive aspect. The participants in the political process might think of themselves as engaging in a competition and judge themselves to be unfairly disadvantaged by their relative lack of input (or their absolute lack of influence). As far as winning the argument goes, or having their preferences or interests fairly registered in the aggregation process, they may be right about that. But if the proper way to assess the decisionmaking procedure is to regard it not as a competition to win the argument or aggregation procedure but as a means to independently correct political decisions, then, on the relevant assumptions, there can be a benefit from inequality. These considerations, and those raised in the previous section, suggest that unfair inequality might be justified, even where that unfairness leaves some people good-specifically worse off than they would otherwise be. As far as children are concerned, it is unfair that differential parental input means that some have better life prospects than others. Yet that unfairness may be justified, all things considered, on prioritarian grounds. It is unfair if those with more resources can use them to improve their children’s chances of success, relative to other people’s 24. Ibid., 182.

488

Ethics

April 2006

children, and it remains so even if that unfairness is justified by the fact that it tends to improve the absolute position of the worse off. It is unfair if some are able to use their wealth to increase their political influence, or devote greater resources to the legal process than their opponents. Yet that unfairness may be justified on the grounds that it serves to increase the probability of the political or legal processes reaching the right decisions.25 In these last cases, what justifies the unfairness is a reframing of the context in noncompetitive terms. VII. REDUCING THE POSITIONAL ASPECT RATHER THAN LEVELING DOWN? The reason to level down with respect to specific goods that we have been discussing applies only insofar as those goods have positional aspects. But leveling down means making some people worse off, with respect to some goods, than they would otherwise be, and we typically have reason not to do that. Rather than leveling them down, would it not be better to reduce (or eliminate altogether) the extent to which goods have positional aspects? In that case, some people could enjoy the benefits of having more rather than less of such goods without the same—perhaps without any—negative impact on the well-being of others. That reducing A’s superior position with respect to X improves B’s absolute position with respect to Y depends on some causal connection between A’s possession of more X than B has and B’s possession of Y. The person with more education has better prospects for income and for accessing interesting and responsible jobs, because there is a causal link between education and labor market prospects. But it is not simply having more education that makes the person’s income prospects better. It is having more education in an environment in which that causal link holds. We could eliminate the causal link between relative education and absolute income by equalizing wage rates. We could reduce the causal link between relative education and absolute chances of getting interesting and responsible jobs by reducing the stigma attached to 25. Estlund disagrees. In his view, “departures from fairness are not always unfair” and “a procedure is unfair only if it wrongly departs from fair procedure” (ibid., 186). He concedes that his politically inegalitarian scheme “is not internally fair since extra influence can be bought by those who can afford it. But it is not unfair, and it is a fair procedure to have if the improvement it brings in the epistemic value of the procedure would be acceptable to all in a fair hypothetical choice procedure” (ibid., 187). For us, it is unfair if those with more resources can use them to increase their chances of winning the election, or the legal case—even if that is justified by the fact that the inequality in input brings with it also greater total input, thereby increasing the likelihood of achieving the independently right result. Fairness would be achieved by distributing the same total amount of resources equally—and the absence of such a distribution is, for us, an unfairness.

Brighouse and Swift

Equality, Priority, and Positional Goods

489

nepotism, by allocating jobs by lottery, or by reforming the job structure to make jobs more equally interesting and responsible. Since education has a nonpositional aspect, and it is other-things-equal desirable for people to get more education regardless of how much others get, one of these alternative strategies might be preferable to leveling down educational provision. Think, alternatively, about the link between income inequality and health. Michael Marmot conjectures that one causal mechanism involved is that higher income correlates with having more control over one’s work and social circumstances and low control triggers stress, which worsens health.26 But, noting that some societies seem to be much less badly affected than others by the link between inequality of income and absolute levels of health, he conjectures that the causal link is weakened or reinforced by other features of the social environment. Societies marked by a certain kind of individualism, and in which there is public endorsement of the idea that the well off have little responsibility to the badly off, are societies in which the link will be more secure than in societies marked by social connectedness and a high sense of mutual responsibility. It may be possible, then, to retain some significant level of income inequality, if that is desirable for, say, the purpose of producing more wealth, while moderating the effect on the health of those with lower incomes, by reforming the public culture or, again, by reforming the structure of occupations so that there are higher levels of control at the lower end of the job ladder. Such reforms would aim to reduce the extent to which inequality of status or income produces stress for those on the low end of the inequality, thereby reducing the negative impact on people’s health that is a direct effect of that stress (and an indirect effect of their relative disadvantage in terms of income or status). Think also of the connection between certain parenting styles and subsequent economic advantage. Annette Lareau argues that middleclass parenting styles are better than working-class ones at preparing children to negotiate their way through the complex institutions they will encounter as adults: “This kind of training developed in Alexander and other middle-class children a sense of entitlement. They felt they had a right to weigh in with an opinion, to make special requests, to pass judgment on others, and to offer advice to adults. They expected to receive attention and to be taken very seriously. It is important to recognize that these advantages and entitlements are historically specific. . . . They are highly effective strategies in the United States today precisely because our society places a premium on assertive, individualized actions executed by persons who command skills in reasoning 26. Marmot, The Status Syndrome, esp. chaps. 6 and 7.

490

Ethics

April 2006

and negotiation.”27 But, she rightly implies, the parenting styles she describes might not confer advantage in a society that prized, for example, modesty, self-effacement, or public spiritedness. Reforming the culture to break the link between the activity and the reward might be preferable to attempting to achieve greater uniformity of parenting styles.28 The final example is, perhaps, more familiar. Inequalities of income and social status have an impact on what we might think of as “social inclusion”; it is easier for wealthier people or for those with higher status to be confident of themselves as full members of society. The poor, especially, but even those merely with lower incomes, are often more marginalized, not just from political decision making but from the cultural and even the physical spaces that constitute “the public sphere.” Leveling income is one strategy for correcting for this, but it is not the only one. We also decommodify certain “inclusion” goods, such as voting, access to politicians, public parks and libraries, schools, and so forth. Our focus in this article is on the prioritarian case for leveling down with respect to positional goods. But leveling down may not always be the right response. Sometimes reducing the positional aspect of a good, by breaking the causal link, may be better. Which kind of strategy is better depends primarily on the relative feasibility of the strategies (breaking the causal link may just not be possible—perhaps only because politically unacceptable—in the circumstances) and on the costs of the policy (breaking the causal link may have costs either to the worse off person or to others that render it undesirable all things considered, just as leveling down may). There are very different ways in which relative position with respect to positional goods affects people’s absolute position with respect to their value or their possession of other goods. The case where goods have competitive value is different from the case where people lack self-respect because they judge their own worth by comparing themselves with others who are better off, in some dimension judged relevant, than themselves. The latter operates via a psychological or subjective mechanism which is, in turn, different from the case where political inequality leads to ill health (because, for example, those with least political influence tend to live in neighborhoods with 27. Annette Lareau, Unequal Childhoods (Berkeley: University of California Press, 2003), 133. 28. It would then become interesting to consider the extent to which middle-class parenting practices are as they are because they have the effect of improving children’s chances of future reward, as opposed to being simply “different,” with future rewards as an unintended by-product. This distinction is central to Swift’s analysis of justifications for sending one’s children to unusually goods schools in How Not to Be a Hypocrite.

Brighouse and Swift

Equality, Priority, and Positional Goods

491

greatest pollution).29 Judgments about the relative merits of leveling down and breaking the link will have to attend carefully to the specificity of these mechanisms, which will vary greatly in the prospects they afford for interruption. VIII. INEQUALITY, INCENTIVES, AND POSITIONAL GOODS The previous section pointed out that, rather than leveling them down, it might be preferable to reduce or even to eliminate the positional aspects of some goods. Let us suppose, however, that, in a particular case, pursuing that option is not feasible or would make things worse than would leveling down. Let us suppose further that attention to the all-things-considered well-being of the worse off gives us reason to permit unfair inequality. Of the various distributions available to us, that which best serves the interests of the worse off (or, perhaps, that which best serves an appropriate weighting of their interests with those of others) is one that permits some to be unfairly advantaged relative to others. These are entirely realistic possibilities, and acknowledging them would seem to lend support to those, hostile to egalitarian thinking, who emphasize the benefits of growth even for the least advantaged. In this final section, we want to relate our discussion of positional goods, and the good-specific egalitarian intuitions they prompt, to the debate about the site of justice and the justifiability of individuals demanding inequality-inducing incentives that has followed in the wake of G. A. Cohen’s critique of Rawls.30 Let us distinguish between (i) a distribution’s being justified all things considered in the circumstances and (ii) a distribution’s being comprehensively justified. The former means that, of the alternative distributions between which one is judging, and having taken into account all the different kinds of value realized by them, the distribution in question is preferable. The latter means that there is no moral basis for criticizing that distribution. Unfair competitions, and policies that make some people worse off than they need to be with respect to specific 29. See J. A. Maantay, “Zoning, Equity and Public Health,” in Health and Social Justice: Politics, Ideology and Inequity in the Distribution of Disease, ed. R. Hofrichter (San Francisco: Wiley, 2003), 228–50, cited in B. Barry, Why Social Justice Matters (Oxford: Polity, 2005). 30. For Cohen’s critique, see G. A. Cohen, If You’re an Egalitarian, How Come You’re So Rich? (Cambridge, MA: Harvard University Press, 2000). For discussion, see L. Murphy, “Institutions and the Demands of Justice,” Philosophy & Public Affairs 27 (1998): 251–91; A. Williams, “Incentives, Inequality and Publicity,” Philosophy & Public Affairs 27 (1998): 226–48; David Estlund, “Liberalism, Equality, and Fraternity in Cohen’s Critique of Rawls,” Journal of Political Philosophy 6 (1998): 99–112; T. Pogge, “On the Site of Distributive Justice: Reflections on Cohen and Murphy,” Philosophy & Public Affairs 29 (2000): 137–69; and A. J. Julius, “Basic Structure and the Value of Equality,” Philosophy & Public Affairs 31 (2003): 321–55.

492

Ethics

April 2006

goods, may be justified in sense i if they make the all-things-considered worst off better off than they would otherwise be. But we can still ask what makes it the case that such competitions and good-specificallynonprioritarian distributions are conducive to the well-being, all things considered in the circumstances, of the least advantaged. Insofar as these factors involve human motivations and actions, it is important that we have the conceptual moral space to criticize them. The distributions may not be justified in sense ii. One influential approach would resist according the moral criticisms we are about to discuss the status of claims about justice. In A Theory of Justice, Rawls says that the basic structure of society is the primary subject of justice and claims that his principles of justice apply specifically to the institutions that constitute that structure. As long as individuals play the game according to the rules set by those institutions, and the institutions themselves are just, there is, on this view, nothing more to be achieved, from the point of view of justice. In terms of our distinction, this line of thought either regards all-things-considered-inthe-circumstances justifications as equivalent to comprehensive justifications or regards the additional considerations invoked in comprehensive justifications as being irrelevant to questions of justice. So it is unnatural (on this line of thought) to regard as unjust the advantages that accrue in competitions we regard as justified-in-the-circumstancesdespite-their-being-unfair. We regard them as justified in the circumstances because they benefit the least advantaged all things considered, but unjust because if the more advantaged individuals had different motivations and pursued different actions it would not be the case that unfair competitions would be needed to produce that benefit.31 We side with G. A. Cohen, in other words, in taking individual motivations and actions to be proper foci of claims about justice. Consider two societies the institutional structures of which both comply with the same distributive principles. In both societies, let us suppose, it is indeed the case that posttax inequalities are such as to maximize the absolute position of the worst off. In one, call it Sweden, those who are lucky enough to have the socially most valued talents are conscientiously egalitarian or prioritarian, so that they will be as productive at higher marginal tax rates as their equivalents in the other (call it the United States). 31. We are talking here specifically about distributive justice, which we equate with fairness. On other conceptions of justice—e.g., where a person is treated unjustly if a right of hers is violated—then there may be justice grounds for permitting distributive injustice. For example, if a parent has the right to pursue her children’s interests to some reasonable extent, justice may require that she be permitted to pursue those interests even at the cost of creating a distributive injustice (or unfairness) between those children and others. On this analysis, it will of course be crucial what weight considerations of distributive justice are to be given relative to these other justice considerations.

Brighouse and Swift

Equality, Priority, and Positional Goods

493

The result, other things being equal, will be less inequality, and a higher absolute position for the worst off, in Sweden than in the United States. Yet a pure basic structure approach does not allow us to say that Sweden is more distributively just than is the United States, because what makes for the differences between them are the motivations and actions of talented individuals, not the principles regulating the structures within which those individuals act.32 The view that there can be no objection to distributions justified on prioritarian grounds is implausible because it lacks the resources to condemn certain mechanisms that nevertheless benefit the least advantaged. Consider the following example: an Oxford college can, for financial reasons, admit only ten students a year and attempts rigorously to identify those most likely to do well in the courses for which they are applying. A wealthy donor offers to fund two extra places for low-income students, on condition that the college creates a third extra place, which she will also fund, for her son. Suppose that the son would not be in the top thirteen on academic merit alone but does have sufficient academic capacity that his presence would not devalue the education of the other twelve students. Assume (implausibly) that there are no effects on third parties. The college may perhaps be justified in accepting the donor’s offer. Nobody is made worse off, and two low-income children are made better off. But it seems hard to deny that there is here some objectionable unfairness. It would be fairer if the same benefit could be achieved without the loss of equality of opportunity. Even those who endorse prioritarian distributive principles should be sensitive to the moral quality of the different kinds of mechanism whereby inequalities benefit the least advantaged. There is something objectionable in the Oxford college example both about the mother seeking to buy advantage for her son and the college selling the advantage. If exactly the same inequality profile had emerged, for whatever reason, without violations of equality of opportunity, the outcome, though identical, 32. Williams, “Incentives, Inequality and Publicity,” rightly objects that Cohen’s extending the site of distributive justice beyond the basic structure to individual motivations and actions deprives it of the valuable property of publicity. We accept both that it is valuable for citizens to live by, and know each other to be living by, public rules and that there will be insuperable difficulties in knowing precisely what justice demands of us as individuals, especially in the absence of assurance that others are acting justly. Neither of these considerations, however, seems to us reason to reject the view that individual motivations and actions are, in principle, susceptible to judgments of justice. For us, publicity is a condition of legitimacy rather than of distributive justice, and even in the absence of the information that would be needed precisely to identify what distributions are just, there are surely cases where an individual has grounds for judging that she is receiving greater (or lesser) benefits than she would were benefits distributed justly.

494

Ethics

April 2006

would be fairer. The mother had the means to fund three places without purchasing advantage for her son. That the transaction between the mother and the college violates fair equality of opportunity renders it morally defective. It may well be justifiable all things considered, on prioritarian grounds.33 But it cannot be comprehensively justified. The circumstances that make it necessary to accept inequality of opportunity as the condition of benefiting the less advantaged are objectionable. The reasons people demand incentives vary widely in moral quality, and this is not the place to attempt an exhaustive discussion of the ways in which such demands may be morally defective. A few pointers may nonetheless be appropriate. One concerns the motivation for the demand. In this example, it seems to us implausible that the mother was acting within the bounds of legitimate partiality, but there will surely be cases where incentive demands do indeed fall within those bounds. Cohen accepts that individuals may properly pursue their self-interest “to some reasonable extent,” and Estlund has persuasively argued that such agent-centered prerogatives must extend to the pursuit of the interests of one’s nearest and dearest—at least in nonideal circumstances. On our analysis, distributions resulting from actions justified by appeal to legitimate partiality may coherently be judged unfair, but it is clearly problematic to judge an agent to be acting within the bounds of legitimate partiality and yet acting unfairly. Another pointer concerns the agent’s knowledge of the justification of the institutional arrangements that allow her to demand incentives. Someone who demands an incentive in the knowledge that the system of incentive payments is designed to satisfy prioritarian principles—or to meet other valued goals—is in a different situation from someone who demands such a payment without regard to whether the system of incentives is so designed. Consider three cases of someone who enjoys more than equal political or legal input: 1. She seeks more than equal input simply to advance her own interests. 2. She seeks more than equal input in the knowledge that her society has devised a system, on Estlundian lines, whereby her additional input will increase the input available to those with least input. 3. She seeks more than equal input in the knowledge both that there are mechanisms in place ensuring that her additional input will increase that available to those with least and that those mechanisms improve the quality of political or legal decisions. Her demands for greater input seem less objectionable as we move 33. G. A. Cohen, “Incentives, Inequality and Community,” in Equal Freedom, ed. Stephen Darwall (Ann Arbor: University of Michigan Press, 1995), 302–3; Estlund, “Liberalism, Equality, and Fraternity in Cohen’s Critique of Rawls.”

Brighouse and Swift

Equality, Priority, and Positional Goods

495

from case 1 to case 3. Case 3 is, for us, though not for Estlund, still objectionable at the bar of fairness. The same total amount of resources, equally distributed, would make the political process fair and improve still further the quality of decisions. Still, it is surely less objectionable than case 2, and even less so than case 1. Our focus in this article has been on positional goods. Good-specific egalitarian intuitions are likely to emerge, we would suggest, as objections to demands for incentives justified on prioritarian grounds, where the goods in question have a manifestly positional aspect. Insofar as goods are positional, one cannot have more than others without negative impact on the value of what others have. Making one’s productive contribution conditional on being allowed to have more political influence than others, or being allowed to hire better lawyers is, we would suggest, more blameworthy than making it conditional on enjoying more consumer goods. Many regard it as less objectionable for people to buy private, and superior, health care, than to buy private, and superior, education for their children. That is because those two goods are widely regarded as having different internal structures. Unlike education, health is not manifestly a competitively positional good. Perhaps, if we knew more about the causal processes by which health inequalities affect people’s chances of success in achieving other valued goods, we might come to see things differently. An analogous point applies in the case of the other, noncompetitive, kind of latently positional goods that we identified in Section III. Wanting to be materially better off than one is and, let us suppose, thereby to be better off in that respect than others, will conventionally seem less objectionable than wanting to be better off with respect to a competitive positional good, such as education. Being better off materially than others does not seem to do them any harm. But what if, in fact, material inequalities—or at least inequalities of a certain extent—negatively affect some people’s happiness or health or self-respect or the degree to which they enjoy the benefits of social inclusion? In that case, it may begin to look as problematic to demand material incentives as to demand reward in the shape of other, more manifestly positional, goods. In these cases, to be sure, the inequalities do not offend against the value of a fair competition, and, for some, they may for that reason be less troubling. But suppose that inequalities in material well-being, or in the distribution of workplace autonomy, negatively affect some people’s health in ways that take them below a threshold of adequacy. Here narrowly—and, we would suggest, more morally urgent—sufficientarian considerations will be enough seriously to problematize the demand for those inequalities.

496

Ethics

April 2006

IX. CONCLUSION Insofar as goods are competitively positional, prioritarians and egalitarians will tend to agree that the distribution should be equal, since there is no scope for increasing the total amount of the good to the benefit of those who have least. Assuming that the supply of top jobs, or places at elite universities, is fixed, at least in the relevant time period, one cannot improve the chances of achieving those goods for poor children without reducing the chances of children born to wealth. Giving poor children a better chance means leveling down with respect to the goods that are competitively valuable in the relevant competitions. Politicians who talk about equality of opportunity seem reluctant to grasp that nettle. Moreover, while some goods have manifest competitive value, for others that aspect of their value is only latent. Any good the differential distribution of which affects people’s chances of succeeding in the competition for other goods is properly conceived as having a competitively positional aspect. Furthermore, there are goods with latent positional value of a noncompetitive kind, where causal mechanisms operate that link one’s place in the distribution of such goods to one’s absolute position with respect to others. In all these cases, leveling down with respect to some goods may mean improving the absolute position of some people with respect to their value, or with respect to other goods. Three considerations, however, challenge the prioritarian case for equality, even for positional goods. First, many positional goods also have nonpositional aspects, which are not in practice detachable from the positional aspects, but the distribution of which also matters, and which mitigates the case for equality. Second, if what we really care about is all-things-considered well-being, then unequal distributions of a particular good that violate prima facie criteria of fairness, and that do not yield more of that good for those who have least of it, may nonetheless be justified, all things considered, on prioritarian grounds. Third, it may be that, within certain limits of inequality and at certain levels of total input, inequality even of competitive goods like political and legal input can bring epistemic gains, improving the quality of decision making. We cannot infer distributive principles for concrete goods of the kind distributed by political policies without knowing how those goods, and, for positional goods, features of their distribution, contribute to the well-being of individuals. We need to be aware of the, often complex, causal mechanisms relating different goods to one another, especially those relating inequalities with respect to some goods to absolutes with respect to others. We need to be aware of them partly because it may be that, rather than leveling down, we could, by interrupting those mechanisms, reduce, or eliminate entirely, the extent to which goods

Brighouse and Swift

Equality, Priority, and Positional Goods

497

are positional. But this is very complicated. Any attempt to develop the arguments of this article in the direction of practical policies will need shorthand rules of thumb, for publicity and other purposes. It may indeed be perverse to favor leveling down with respect to a suitably sophisticated metric of advantage. If unfair inequality is the price of benefiting the worst off, then that price may indeed be worth paying. But that leaves much of importance unsaid. A comprehensive justification of that inequality would include a justification of the motivations that make it the case that we have to choose between inequality justified on prioritarian grounds and equality at a lower level. Some such motivations are surely justified, but some are not. The extent of our moral disapproval depends on the structure of the good involved, whether the motivation falls within the scope of legitimate partiality, and the agent’s awareness of both the institutional context and the overall justification of institutions and rules. Demands for inequality in terms of goods with positional aspects are, we suggest, distinctively problematic, because being better off than others with respect to such goods has some adverse impact on the absolute position of those others.